How Discovery Works

The modern era of vast, collaborative, and expensive instruments — the great space observatories, the kilometre-scale gravitational-wave detectors, the all-sky surveys — that no single scientist could build or fund, run instead by international collaborations of thousands.

The recurring pattern by which a new instrument opens a new window on the universe and a wave of discovery follows — the telescope, the spectroscope, the photographic plate, the radio dish, the gravitational-wave detector each launching a new astronomy.

The systematic ways instruments and selection effects skew what astronomers find — brighter objects being easier to see, for instance, biasing a sample. Recognising and correcting these biases is as much a part of a result as the measurement itself.

The deep changes in how a field sees the world, as described by the philosopher Thomas Kuhn — moments when one framework of assumptions gives way to another, as when the Earth-centred cosmos was replaced by a Sun-centred one.

The episodes in which accumulated anomalies — observations that will not fit the old picture — build until the framework breaks and a new one takes its place. Astronomy has driven several of the largest revolutions in all of science.

The transformation of astronomy into a data science, in which surveys now produce far more data than any team could inspect by hand, and machine learning is used to classify, discover, and even find what no one thought to look for.

The cycle of observation, hypothesis, prediction, and test that turned the study of the heavens from stargazing and myth into a science — and the reason astronomy's claims can be checked, challenged, and improved rather than merely believed.

The interplay between prediction and measurement that drives astronomy forward — sometimes theory leads and observation confirms (the prediction of Neptune, of gravitational waves), and sometimes observation surprises and theory must catch up (the accelerating universe).